Mirra
Your AI use patterns and how you solve problems can help you find a cofounder.
Mirra finds and uses evidence to analyze how you and a partner handle hard moments, using your own AI chat history as the evidence.
The problem
Most cofounder compatibility advice is vibes.
You don't need another personality quiz. You need real evidence of how you each actually behave under disagreement, challenge, and risk — pulled from AI conversations you've already had, not a survey you fill out in the moment.
How the partner path works
01
Pick a question.
Choose from 4 preloaded questions built for cofounders.
02
Mine your own AI history.
You both run the same Mirra prompt in your own AI and paste back what it finds.
03
Resolve, then discuss.
Mirra compares both sides and shows where you align, where you diverge, and what to talk about next. Neither of you sees it until both sides are in.
Two ways in
Start on your own, or start together.
Both paths run the same prompt against the same kind of evidence, and both let you review everything before it goes anywhere. The only difference is whether a second person is involved.
On your own
See what your own history says.
Run the prompt against your own AI conversations and get back the specific moments where you handled disagreement, changed your mind, or weighed risk. Review every quote it pulls, move anything mis-tagged, remove anything that isn't you.
Needs nothing but your own chat history.
With a partner
Compare how you each handle hard moments.
Pick a question and send it to someone you're thinking of building with. You each run the same prompt against your own history, and Mirra shows where you align, where you diverge, and what to talk about next — with receipts.
Needs a second person to paste their side before it resolves.
Where the evidence comes from
The prompt draws on conversations you've already had, with your AI's memory switched on — so it surfaces how you actually reasoned in the moment, not answers you'd polish for a quiz. You review everything it pulls before it's ever shared.
What comes back
A resolution shows where you align, where you diverge, and what to talk about next — each claim backed by receipts from your real conversations. Each side sees a paraphrased view; your own near-verbatim quotes only ever expand for you.
What we can and can't see
Mirra is not a personality test or a diagnostic, and it can't predict how well two people will actually work together. It's a private way to explore your AI conversations with someone else. This is an early alpha: raw inputs auto-delete after 7 days, resolutions after 30. Either way, don't paste anything you wouldn't want a stranger to see.